Abstract: The human microbiome - the complex ensemble of microorganisms that populate the human body - has a tremendous impact on our health. World-wide research initiatives now provide preliminary insights into the previously uncharted composition of the microbiome, and reveal marked compositional changes associated with a wide range of diseases and host phenotypes. The gut microbiome, especially, plays a key role in many essential processes and actively contributes to various disease states, including obesity, diabetes, inflammatory bowel disease, cardiovascular diseases, and neurological disorders. Specifically, recent studies have demonstrated that transferring a donor microbiota into a recipient can induce various donor phenotypes or prompt the recovery of a sick recipient. These findings suggest a promising therapeutic avenue via directed manipulation of the gut microbiome. Such clinical interventions could target, for example, a microbiome whose composition typifies a certain disease state, such as diabetes, and promote a compositional shift into a healthy configuration. Alternatively, individuals could be colonized with a new designer microbiome with some preferred metabolic capacities, allowing, for example, populations of undernourished children to harvest more energy from a limited diet. However, to allow microbiome-based therapy to move forward and to realize its full potential, a comprehensive computational toolkit is required for directing such manipulations and proposing promising intervention routes. In this project, we will accordingly develop a computational framework for designing microbiome manipulation targeted at a specific set of desired metabolic goals. This framework will encompass two main components: A system-level in-silico model of microbiome metabolism, capable of successfully predicting the metabolic activity of a given microbial consortia in the gut environment, and an optimization module which will be used to search the space of possible microbiome compositions for those that most closely match the required metabolic goals. We will use multiple metabolic modeling and analysis approaches, and present novel computational methods for studying complex multi-species communities. Such system-level computational methods proved extremely powerful and effective in studying and engineering single species metabolism, but have not yet been applied to study community-wide metabolism. Computational techniques will also be developed to account for varying species abundances and to assess the resilience of designed microbiomes. This proposed cross-disciplinary research integrates computational systems biology, in-silico models, and complex networks analysis, with genomic and metagenomic data, taking a first crucial step in the construction of a computational framework for guiding and informing microbiome-based clinical interventions. This project represents a major leap forward in the study of the human microbiome and in providing computational tools for harnessing this newly gained knowledge directly for promoting human health. Public Health Relevance: The human microbiome has a tremendous impact on our health and has been associated with various disease states ranging from obesity and diabetes to cardiovascular diseases and neurological disorders. Microbiome manipulation, either via targeted specific intervention or via whole microbiome transplantation, is an exciting clinical frontier with numerous promising medical applications. In this project, we will develop a novel computational framework, integrating a predictive model of the microbiome and its impact on the host with optimization techniques, for designing such manipulations and for informing clinical intervention efforts.